We finetuned gptj-6b on Code-Alpaca-Instruct Dataset (vicgalle/alpaca-gpt4) for 10 epochs or ~ 50,000 steps using MonsterAPI no-code LLM finetuner.
This dataset is vicgalle/alpaca-gpt4 unfiltered,
The finetuning session got completed in 7 hours and costed us only $25 for the entire finetuning run!
Hyperparameters & Run details:
- Model Path: vicgalle/alpaca-gpt4
- Dataset: vicgalle/alpaca-gpt4
- Learning rate: 0.0003
- Number of epochs: 5
- Data split: Training: 90% / Validation: 10%
- Gradient accumulation steps: 1
license: apache-2.0
- Downloads last month
- 1
Unable to determine this model’s pipeline type. Check the
docs
.